Can I use this with an existing codebase?

Yes, it includes a 'Map Codebase' command that analyzes your current stack, architecture, and conventions to ensure the AI understands your project structure before starting work.

Does it support git workflows?

Yes, the skill enforces atomic git commits linked to specific tasks and verification steps, ensuring a clean and traceable version control history.

What is 'auto-sizing' in this skill?

Auto-sizing dynamically adjusts the planning depth—skipping design or task phases for small fixes while requiring full documentation and architectural analysis for complex features.

How does it handle long development sessions?

It uses a persistent STATE.md file to track decisions, blockers, and progress, allowing you to pause and resume work seamlessly without losing context.

Spec-Driven Development (TLC)

Name: Spec-Driven Development (TLC)
Author: tech-leads-club

bytech-leads-club

•

1,856

•

Productivity & Workflow

Implements an adaptive 4-phase workflow—Specify, Design, Task, and Execute—to ensure high-precision project planning and implementation.

The Spec-Driven Development skill provides a disciplined framework for managing the software development lifecycle within AI coding environments. By utilizing an adaptive 'auto-sizing' principle, it intelligently adjusts its depth based on task complexity, ranging from an express 'Quick mode' for bug fixes to a full-scale architectural design for complex features. The skill maintains project integrity through a structured .specs/ directory, ensuring requirement traceability, persistent session memory, and atomic git commits, making it the ideal tool for both starting greenfield projects and mapping existing brownfield codebases.

Key Features

011,856 GitHub stars

02Persistent session memory (STATE.md) to track decisions, blockers, and deferred ideas across AI sessions.

03Adaptive 4-phase workflow (Specify, Design, Tasks, Execute) that auto-sizes based on complexity.

04Requirement traceability with unique IDs linked to atomic tasks and automated verification criteria.

05Comprehensive brownfield mapping to document existing codebases across 7 key architectural dimensions.

06Integrated 'Quick mode' for small tweaks and bug fixes to maintain speed without sacrificing documentation.

Use Cases

01Initializing a new project with a vision, roadmap, and structured goal alignment.

02Executing complex features with granular task breakdowns and atomic git commits for cleaner reviews.

03Onboarding an AI agent to an existing codebase by generating stack, architecture, and convention docs.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tech-leads-club/agent-skills (development)

For use in Claude.ai and ChatGPT

Key Features

011,856 GitHub stars

02Persistent session memory (STATE.md) to track decisions, blockers, and deferred ideas across AI sessions.

03Adaptive 4-phase workflow (Specify, Design, Tasks, Execute) that auto-sizes based on complexity.

04Requirement traceability with unique IDs linked to atomic tasks and automated verification criteria.

05Comprehensive brownfield mapping to document existing codebases across 7 key architectural dimensions.

06Integrated 'Quick mode' for small tweaks and bug fixes to maintain speed without sacrificing documentation.

Use Cases

01Initializing a new project with a vision, roadmap, and structured goal alignment.

02Executing complex features with granular task breakdowns and atomic git commits for cleaner reviews.

03Onboarding an AI agent to an existing codebase by generating stack, architecture, and convention docs.